From ivan at azulsystems.com  Wed May  6 13:01:30 2015
From: ivan at azulsystems.com (Ivan Krylov)
Date: Wed, 6 May 2015 16:01:30 +0300
Subject: Build failure on zero/openjdk9
Message-ID: <554A10AA.5040801@azulsystems.com>

Hi folks,

Today I found a small error in the printf parameters that prevents 
openjdk9 to build in a zero configuration.
Who owns hotspot/src/cpu/zero/vm/frame_zero.cpp ?

This is the build error:
/home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp: In member 
function ?void frame::zero_print_on_error(int, outputStream*, char*, 
int) const?:
/home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp:216:73: 
error: format ?%lx? expects argument of type ?long unsigned int?, but 
argument 3 has type ?intptr_t* {aka long int*}? [-Werror=format=]
      st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
^
cc1plus: all warnings being treated as errors

---

The patch is trivial (I am the OCA signatory as ikrylov)

ivan:~/openjdk9/hotspot$ hg diff
diff -r 86ab44ce262e src/cpu/zero/vm/frame_zero.cpp
--- a/src/cpu/zero/vm/frame_zero.cpp    Thu Apr 30 13:02:03 2015 -0700
+++ b/src/cpu/zero/vm/frame_zero.cpp    Wed May 06 15:55:10 2015 +0300
@@ -213,7 +213,7 @@
      valuebuf[buflen - 1] = '\0';

      // Print the result
-    st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
+    st->print_cr(" " PTR_FORMAT ": %-21s = %s", p2i(addr), fieldbuf, 
valuebuf);
    }
  }

Can someone push the change or do I need to file the bug/go through code 
review process?

Thanks,

Ivan




From volker.simonis at gmail.com  Thu May  7 09:46:06 2015
From: volker.simonis at gmail.com (Volker Simonis)
Date: Thu, 7 May 2015 11:46:06 +0200
Subject: Build failure on zero/openjdk9
In-Reply-To: <554A10AA.5040801@azulsystems.com>
References: <554A10AA.5040801@azulsystems.com>
Message-ID: <CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>

Hi Ivan,

the change looks good and I can sponsor it. But you always need a bug
ID so yes, please open a bug for it.

I also put Severin from RedHat on CC to hear his opinion because as
far as I know he's currently maintaining the Zero port.

Regards,
Volker


On Wed, May 6, 2015 at 3:01 PM, Ivan Krylov <ivan at azulsystems.com> wrote:
> Hi folks,
>
> Today I found a small error in the printf parameters that prevents openjdk9
> to build in a zero configuration.
> Who owns hotspot/src/cpu/zero/vm/frame_zero.cpp ?
>
> This is the build error:
> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp: In member
> function ?void frame::zero_print_on_error(int, outputStream*, char*, int)
> const?:
> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp:216:73: error:
> format ?%lx? expects argument of type ?long unsigned int?, but argument 3
> has type ?intptr_t* {aka long int*}? [-Werror=format=]
>      st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
> ^
> cc1plus: all warnings being treated as errors
>
> ---
>
> The patch is trivial (I am the OCA signatory as ikrylov)
>
> ivan:~/openjdk9/hotspot$ hg diff
> diff -r 86ab44ce262e src/cpu/zero/vm/frame_zero.cpp
> --- a/src/cpu/zero/vm/frame_zero.cpp    Thu Apr 30 13:02:03 2015 -0700
> +++ b/src/cpu/zero/vm/frame_zero.cpp    Wed May 06 15:55:10 2015 +0300
> @@ -213,7 +213,7 @@
>      valuebuf[buflen - 1] = '\0';
>
>      // Print the result
> -    st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
> +    st->print_cr(" " PTR_FORMAT ": %-21s = %s", p2i(addr), fieldbuf,
> valuebuf);
>    }
>  }
>
> Can someone push the change or do I need to file the bug/go through code
> review process?
>
> Thanks,
>
> Ivan
>
>
>

From ivan at azulsystems.com  Thu May  7 11:55:36 2015
From: ivan at azulsystems.com (Ivan Krylov)
Date: Thu, 7 May 2015 14:55:36 +0300
Subject: Build failure on zero/openjdk9
In-Reply-To: <CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>
References: <554A10AA.5040801@azulsystems.com>
	<CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>
Message-ID: <554B52B8.9020405@azulsystems.com>

Ok, I have filed a new bug https://bugs.openjdk.java.net/browse/JDK-8079594
Synopsis: Build failure for Zero target
The patch can be viewed at http://cr.openjdk.java.net/~ikrylov/8079594/

Thanks Volker,
Ivan

On 07/05/2015 12:46, Volker Simonis wrote:
> Hi Ivan,
>
> the change looks good and I can sponsor it. But you always need a bug
> ID so yes, please open a bug for it.
>
> I also put Severin from RedHat on CC to hear his opinion because as
> far as I know he's currently maintaining the Zero port.
>
> Regards,
> Volker
>
>
> On Wed, May 6, 2015 at 3:01 PM, Ivan Krylov <ivan at azulsystems.com> wrote:
>> Hi folks,
>>
>> Today I found a small error in the printf parameters that prevents openjdk9
>> to build in a zero configuration.
>> Who owns hotspot/src/cpu/zero/vm/frame_zero.cpp ?
>>
>> This is the build error:
>> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp: In member
>> function ?void frame::zero_print_on_error(int, outputStream*, char*, int)
>> const?:
>> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp:216:73: error:
>> format ?%lx? expects argument of type ?long unsigned int?, but argument 3
>> has type ?intptr_t* {aka long int*}? [-Werror=format=]
>>       st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
>> ^
>> cc1plus: all warnings being treated as errors
>>
>> ---
>>
>> The patch is trivial (I am the OCA signatory as ikrylov)
>>
>> ivan:~/openjdk9/hotspot$ hg diff
>> diff -r 86ab44ce262e src/cpu/zero/vm/frame_zero.cpp
>> --- a/src/cpu/zero/vm/frame_zero.cpp    Thu Apr 30 13:02:03 2015 -0700
>> +++ b/src/cpu/zero/vm/frame_zero.cpp    Wed May 06 15:55:10 2015 +0300
>> @@ -213,7 +213,7 @@
>>       valuebuf[buflen - 1] = '\0';
>>
>>       // Print the result
>> -    st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
>> +    st->print_cr(" " PTR_FORMAT ": %-21s = %s", p2i(addr), fieldbuf,
>> valuebuf);
>>     }
>>   }
>>
>> Can someone push the change or do I need to file the bug/go through code
>> review process?
>>
>> Thanks,
>>
>> Ivan
>>
>>
>>


From gnu.andrew at redhat.com  Mon May 11 02:29:16 2015
From: gnu.andrew at redhat.com (Andrew Hughes)
Date: Sun, 10 May 2015 22:29:16 -0400 (EDT)
Subject: Build failure on zero/openjdk9
In-Reply-To: <CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>
References: <554A10AA.5040801@azulsystems.com>
	<CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>
Message-ID: <905884312.14281080.1431311356403.JavaMail.zimbra@redhat.com>

----- Original Message -----
> Hi Ivan,
> 
> the change looks good and I can sponsor it. But you always need a bug
> ID so yes, please open a bug for it.
> 
> I also put Severin from RedHat on CC to hear his opinion because as
> far as I know he's currently maintaining the Zero port.
> 

It's "Red Hat", not RedHat :)

This looks fine to me, though I'm curious as to why it's causing issues
now. This line of code is present all the way back to OpenJDK 6 and comes
from the original Zero integration.

changeset:   1010:354d3184f6b2
user:        never
date:        Tue Oct 13 12:04:21 2009 -0700
summary:     6890308: integrate zero assembler hotspot changes

A new version of GCC? New flags being enabled by default?

> Regards,
> Volker
> 
> 
> On Wed, May 6, 2015 at 3:01 PM, Ivan Krylov <ivan at azulsystems.com> wrote:
> > Hi folks,
> >
> > Today I found a small error in the printf parameters that prevents openjdk9
> > to build in a zero configuration.
> > Who owns hotspot/src/cpu/zero/vm/frame_zero.cpp ?
> >
> > This is the build error:
> > /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp: In member
> > function ?void frame::zero_print_on_error(int, outputStream*, char*, int)
> > const?:
> > /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp:216:73: error:
> > format ?%lx? expects argument of type ?long unsigned int?, but argument 3
> > has type ?intptr_t* {aka long int*}? [-Werror=format=]
> >      st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
> > ^
> > cc1plus: all warnings being treated as errors
> >
> > ---
> >
> > The patch is trivial (I am the OCA signatory as ikrylov)
> >
> > ivan:~/openjdk9/hotspot$ hg diff
> > diff -r 86ab44ce262e src/cpu/zero/vm/frame_zero.cpp
> > --- a/src/cpu/zero/vm/frame_zero.cpp    Thu Apr 30 13:02:03 2015 -0700
> > +++ b/src/cpu/zero/vm/frame_zero.cpp    Wed May 06 15:55:10 2015 +0300
> > @@ -213,7 +213,7 @@
> >      valuebuf[buflen - 1] = '\0';
> >
> >      // Print the result
> > -    st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
> > +    st->print_cr(" " PTR_FORMAT ": %-21s = %s", p2i(addr), fieldbuf,
> > valuebuf);
> >    }
> >  }
> >
> > Can someone push the change or do I need to file the bug/go through code
> > review process?
> >
> > Thanks,
> >
> > Ivan
> >
> >
> >
> 

-- 
Andrew :)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222

PGP Key: rsa4096/248BDC07 (hkp://keys.gnupg.net)
Fingerprint = EC5A 1F5E C0AD 1D15 8F1F  8F91 3B96 A578 248B DC07


From ivan at azulsystems.com  Tue May 12 06:41:30 2015
From: ivan at azulsystems.com (Ivan Krylov)
Date: Tue, 12 May 2015 09:41:30 +0300
Subject: Build failure on zero/openjdk9
In-Reply-To: <905884312.14281080.1431311356403.JavaMail.zimbra@redhat.com>
References: <554A10AA.5040801@azulsystems.com>
	<CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>
	<905884312.14281080.1431311356403.JavaMail.zimbra@redhat.com>
Message-ID: <5551A09A.7090300@azulsystems.com>

Hi Andrew,

Typically such errors pop up when a compiler version being updated.
I am on Ubuntu 14.04.2 LTS with gcc 4.8.2
I assume that compiler got a bit more strict about implicit 
signed/unsigned conversions.
I think that the stock hotspot always had treat-warning-as-error flag 
on, but I am not sure if zero was always that way too.

Thanks,

Ivan


On 11/05/2015 05:29, Andrew Hughes wrote:
> ----- Original Message -----
>> Hi Ivan,
>>
>> the change looks good and I can sponsor it. But you always need a bug
>> ID so yes, please open a bug for it.
>>
>> I also put Severin from RedHat on CC to hear his opinion because as
>> far as I know he's currently maintaining the Zero port.
>>
> It's "Red Hat", not RedHat :)
>
> This looks fine to me, though I'm curious as to why it's causing issues
> now. This line of code is present all the way back to OpenJDK 6 and comes
> from the original Zero integration.
>
> changeset:   1010:354d3184f6b2
> user:        never
> date:        Tue Oct 13 12:04:21 2009 -0700
> summary:     6890308: integrate zero assembler hotspot changes
>
> A new version of GCC? New flags being enabled by default?
>
>> Regards,
>> Volker
>>
>>
>> On Wed, May 6, 2015 at 3:01 PM, Ivan Krylov <ivan at azulsystems.com> wrote:
>>> Hi folks,
>>>
>>> Today I found a small error in the printf parameters that prevents openjdk9
>>> to build in a zero configuration.
>>> Who owns hotspot/src/cpu/zero/vm/frame_zero.cpp ?
>>>
>>> This is the build error:
>>> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp: In member
>>> function ?void frame::zero_print_on_error(int, outputStream*, char*, int)
>>> const?:
>>> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp:216:73: error:
>>> format ?%lx? expects argument of type ?long unsigned int?, but argument 3
>>> has type ?intptr_t* {aka long int*}? [-Werror=format=]
>>>       st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
>>> ^
>>> cc1plus: all warnings being treated as errors
>>>
>>> ---
>>>
>>> The patch is trivial (I am the OCA signatory as ikrylov)
>>>
>>> ivan:~/openjdk9/hotspot$ hg diff
>>> diff -r 86ab44ce262e src/cpu/zero/vm/frame_zero.cpp
>>> --- a/src/cpu/zero/vm/frame_zero.cpp    Thu Apr 30 13:02:03 2015 -0700
>>> +++ b/src/cpu/zero/vm/frame_zero.cpp    Wed May 06 15:55:10 2015 +0300
>>> @@ -213,7 +213,7 @@
>>>       valuebuf[buflen - 1] = '\0';
>>>
>>>       // Print the result
>>> -    st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf, valuebuf);
>>> +    st->print_cr(" " PTR_FORMAT ": %-21s = %s", p2i(addr), fieldbuf,
>>> valuebuf);
>>>     }
>>>   }
>>>
>>> Can someone push the change or do I need to file the bug/go through code
>>> review process?
>>>
>>> Thanks,
>>>
>>> Ivan
>>>
>>>
>>>


From volker.simonis at gmail.com  Wed May 13 09:07:11 2015
From: volker.simonis at gmail.com (Volker Simonis)
Date: Wed, 13 May 2015 11:07:11 +0200
Subject: Build failure on zero/openjdk9
In-Reply-To: <5551A09A.7090300@azulsystems.com>
References: <554A10AA.5040801@azulsystems.com>
	<CA+3eh1014GuZw7N21HS+LFv56Ff9ByVqkhu6DTuzWQqxTgW03Q@mail.gmail.com>
	<905884312.14281080.1431311356403.JavaMail.zimbra@redhat.com>
	<5551A09A.7090300@azulsystems.com>
Message-ID: <CA+3eh13js0nREfU9FnjHd8Ni+M1OaRV3cZd5mRuBjVbTZ8t_Zw@mail.gmail.com>

Hi,

I just saw that this was already fixed as part of "8075967: Zero
interpreter asserts for SafeFetch<32,N> calls in ObjectMonitor" [1] in
the jdk9/hs-rt/hotspot repository [2]. You just have to wait until it
arrives in jdk9/dev/hotspot.

So I'll close this bug as duplicate.

Regards,
Volker

[1] https://bugs.openjdk.java.net/browse/JDK-8075967
[2] http://hg.openjdk.java.net/jdk9/hs-rt/hotspot/rev/2206bbeb3185


On Tue, May 12, 2015 at 8:41 AM, Ivan Krylov <ivan at azulsystems.com> wrote:
> Hi Andrew,
>
> Typically such errors pop up when a compiler version being updated.
> I am on Ubuntu 14.04.2 LTS with gcc 4.8.2
> I assume that compiler got a bit more strict about implicit signed/unsigned
> conversions.
> I think that the stock hotspot always had treat-warning-as-error flag on,
> but I am not sure if zero was always that way too.
>
> Thanks,
>
> Ivan
>
>
>
> On 11/05/2015 05:29, Andrew Hughes wrote:
>>
>> ----- Original Message -----
>>>
>>> Hi Ivan,
>>>
>>> the change looks good and I can sponsor it. But you always need a bug
>>> ID so yes, please open a bug for it.
>>>
>>> I also put Severin from RedHat on CC to hear his opinion because as
>>> far as I know he's currently maintaining the Zero port.
>>>
>> It's "Red Hat", not RedHat :)
>>
>> This looks fine to me, though I'm curious as to why it's causing issues
>> now. This line of code is present all the way back to OpenJDK 6 and comes
>> from the original Zero integration.
>>
>> changeset:   1010:354d3184f6b2
>> user:        never
>> date:        Tue Oct 13 12:04:21 2009 -0700
>> summary:     6890308: integrate zero assembler hotspot changes
>>
>> A new version of GCC? New flags being enabled by default?
>>
>>> Regards,
>>> Volker
>>>
>>>
>>> On Wed, May 6, 2015 at 3:01 PM, Ivan Krylov <ivan at azulsystems.com> wrote:
>>>>
>>>> Hi folks,
>>>>
>>>> Today I found a small error in the printf parameters that prevents
>>>> openjdk9
>>>> to build in a zero configuration.
>>>> Who owns hotspot/src/cpu/zero/vm/frame_zero.cpp ?
>>>>
>>>> This is the build error:
>>>> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp: In member
>>>> function ?void frame::zero_print_on_error(int, outputStream*, char*,
>>>> int)
>>>> const?:
>>>> /home/ivan/openjdk9/hotspot/src/cpu/zero/vm/frame_zero.cpp:216:73:
>>>> error:
>>>> format ?%lx? expects argument of type ?long unsigned int?, but argument
>>>> 3
>>>> has type ?intptr_t* {aka long int*}? [-Werror=format=]
>>>>       st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf,
>>>> valuebuf);
>>>> ^
>>>> cc1plus: all warnings being treated as errors
>>>>
>>>> ---
>>>>
>>>> The patch is trivial (I am the OCA signatory as ikrylov)
>>>>
>>>> ivan:~/openjdk9/hotspot$ hg diff
>>>> diff -r 86ab44ce262e src/cpu/zero/vm/frame_zero.cpp
>>>> --- a/src/cpu/zero/vm/frame_zero.cpp    Thu Apr 30 13:02:03 2015 -0700
>>>> +++ b/src/cpu/zero/vm/frame_zero.cpp    Wed May 06 15:55:10 2015 +0300
>>>> @@ -213,7 +213,7 @@
>>>>       valuebuf[buflen - 1] = '\0';
>>>>
>>>>       // Print the result
>>>> -    st->print_cr(" " PTR_FORMAT ": %-21s = %s", addr, fieldbuf,
>>>> valuebuf);
>>>> +    st->print_cr(" " PTR_FORMAT ": %-21s = %s", p2i(addr), fieldbuf,
>>>> valuebuf);
>>>>     }
>>>>   }
>>>>
>>>> Can someone push the change or do I need to file the bug/go through code
>>>> review process?
>>>>
>>>> Thanks,
>>>>
>>>> Ivan
>>>>
>>>>
>>>>
>